物理化学学报(Wuli Huaxue Xuebao)
Acta Phys. -Chim. Sin. 2014, 30 (11), 2000-2008
2000
[Article]
doi: 10.3866/PKU.WHXB201408291
November
www.whxb.pku.edu.cn
单壁碳纳米管上毒性苯气体净化的分子模拟
彭
璇*
(北京化工大学信息科学与技术学院, 北京 100029)
摘要:
采用巨正则系综蒙特卡罗(GCMC)方法研究了空气中微量苯组分在单臂碳纳米管(SWNTs)上的吸附净
化. 模拟表明, 具有较大孔径的(20,20)纳米管比较适合吸附纯苯蒸汽, 而对于移除空气中的毒性苯物质, 苯的
吸附选择性分别在(12,12)纳米管及 4.0 MPa 时和(18,18)纳米管及 0.1 MPa 时出现最小值和最大值. 为了解释
这一异常行为, 我们进一步分析了 N2-O2-C6H6 混合物的局部密度分布、吸附分子构型和概率密度分布, 发现(18,
18)纳米管内外完全被苯分子占据, 而对于(12,12)纳米管, 由于存在更强的吸附质-吸附剂相互作用, 空气分子
更倾向于吸附在管与管之间的间隙. 此外, 吸附分子的空间有序参数表明大多数苯分子采取“平躺”在纳米管表
面的定位, 而线性的 N2 和 O2 分子则多数平行于孔轴方向. 最后研究了温度和苯分子主体相浓度对分离效果的
影响. 我们发现较大孔中的选择性随着温度的增加比小孔下降更加明显. 与此对比, 主体相苯浓度对小孔中的
选择性起到更加重要的作用.
关键词:
巨正则系综蒙特卡罗; 吸附; 分离; 苯; 空气; 碳纳米管
中图分类号:
O647
Molecular Simulations of the Purification of Toxic Benzene Gas on
Single-Walled Carbon Nanotubes
PENG Xuan*
(College of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, P. R. China)
Abstract: Grand canonical ensemble Monte Carlo (GCMC) simulations were performed to investigate the
purification of benzene from air by single-walled carbon nanotubes (SWNTs). It was found that (20,20) SWNT
with a large diameter is suitable to adsorb pure benzene. For the removal of benzene in air, the minimum and
maximum selectivities were observed for the (12,12) SWNT at 4.0 MPa and the (18,18) SWNT at 0.1 MPa,
respectively. To obtain deep insight into the unusual behavior, we analyzed the local density profiles, snapshots,
and probability profiles of N2-O2-C6H6 mixtures. The results showed that the (18,18) SWNT was entirely occupied
by benzene molecules, while, for the (12,12) SWNT, N2 and O2 were prone to appear in the interstices between
tubes, instead of inside tubes, because of stronger adsorbate-adsorbent interactions. Additionally, we calculated
the orientation order parameters of the adsorbates. The results suggested that benzene molecules prefer lying
nearly flat on the pore surface, while N2 and O2 molecules orient parallel to the pore axis. Finally, the effects of
temperature and concentration on the selectivity of benzene were investigated. We found that with increasing
temperature the selectivity in large pores decreased more evidently than that in small pores. By contrast, the
concentration plays a more important role in affecting the selectivity in small pores.
Key Words:
Grand canonical ensemble Monte Carlo; Adsorption;
Carbon nanotube
Separation;
Benzene; Air;
Received: June 4, 2014; Revised: August 29, 2014; Published on Web: August 29, 2014.
∗
Corresponding author. Email: [email protected], [email protected]; Tel: +86-10-64430917.
The project was supported by the Open Project of State Key Laboratory of Chemical Engineering, China (SKL-Che-12C01).
化学工程联合国家重点实验室开放课题(SKL-Che-12C01)资助
© Editorial office of Acta Physico-Chimica Sinica
No.11 PENG Xuan: Molecular Simulations of the Purification of Toxic Benzene Gas on Single-Walled Carbon Nanotubes
1
Introduction
In the production process of spraying, printing, leather processing that use large amounts of organic solvents, there will
escape waste gas containing the substances of benzene, toluene,
and xylene. It is reported that a concentration in the air as low as
l00×10- 6 is considered to constitute a definite hazard.1 Directly
releasing the waste gas to atmosphere is closely related to environmental pollution and health safety. Consequently, control of
their emissions has become increasingly urgent and crucial importance on global atmospheric chemistry and quality of life. The
treatment of the waste gas includes catalytic combustion method,
absorption method, biochemical method, and adsorption method.2
At present, a more consistent view is to use the adsorption method.
Up to now, adsorptions of benzene on the adsorbent materials
such as activated carbons, zeolites, and metal organic frameworks
(MOFs)3,4 have been widely studied. Do and Do5 investigated the
effect of intermolecular potential models on the adsorption of
benzene on graphitized thermal carbon black. Coasne et al.6 investigated the structure and dynamics of benzene confined in the
MCM-41 silica nanopores with different diameters by means of
grand canonical Monte Carlo (GCMC) and molecular dynamics
simulations. Similarly, Jousse et al.7 performed a variety of force
field based simulations to study the location and diffusion of
benzene adsorbed in a model zeolite HY with the Si/Al mole ratio
of 2.43. Amirjalayer et al.8 performed molecular dynamic simulations to investigate the effect of lattice dynamics on the benzene
diffusion in MOFs. Besides the adsorbents mentioned, carbon
nanotube is also an extremely important porous material. Since
Iijima9 discovered the carbon nanotubes in 1991, there appear
numerous experimental and computational studies on the application of carbon nanotubes, especially for single-walled carbon
nanotubes (SWNTs).10- 12 For example, Cinke et al.10 discovered
that the SWNTs can adsorb nearly twice the volume of CO2,
compared to activated carbon. Wang et al.11 investigated the selective adsorption of H2S and SO2 from natural gas and flue gas
via GCMC method and found that the SWNTs with optimized
pore sizes were particularly suitable to capture trace sulfur gases.
Huang et al.12 simulated the adsorption separation of equimolar
CH4- CO2 mixture and found that the SWNTs have a higher selectivity of CO2 than activated carbon, zeolites 13X, and MOFs.
It seems that SWNTs could be very promising to gas storage
and separation for their unique pore structure, modifiable surface
chemistry, and controlled pore sizes. However, to our knowledge,
the data of the adsorption efficiency and optimum conditions for
single and mixed benzene pollutants are still very rare, and these
data are the extremely important parameters of design and operation for the pressure swing adsorption (PSA) process. In this
paper, by performing molecular simulations for the ternary
mixture of C6H6, N2, and O2, we systematically investigate the
effects of temperature, pressure, pore size, and benzene concentration on the adsorption behavior of SWNTs. It is expected that
this work can provide theoretical basis for the carbon nanotube
adsorption treatment of waste gas containing low concentration of
2001
benzene in engineering design and practice, as well as the control
of reasonable process parameters.
2
Molecular simulations
In our simulations, air is regarded as a mixture of N2 and O2
with an approximate mole ratio of 79:21. All-atom forcefield is
used to represent the three fluid molecules. Benzene is described
by a nine-site potential,13 where the positive charge of +8.130e is
located at the ring center, and two negative charges of -4.065e
represent the π-electron clouds above and below the benzene ring.
The negative charges are located at a position z=±0.4 nm along the
symmetry axis normal to the benzene ring. Each CH group is
represented as a single interaction site that interacts with a
Lennard-Jones (LJ) potential. For oxygen molecules, we used the
two-site LJ intermolecular potential model from Perng et al.,14
where the two sites are centered on the nuclear positions and their
LJ parameters are determined by fitting experimental liquid phase
diagram to the molecular dynamics simulations. Similarly, the
potential model of N2 is taken from the three-site TraPPE force
field optimized for VLE calculations.15 Each nitrogen atom is
modeled by a LJ site separated by the experimental bond length
of 0.11 nm. A point charge of -0.482e is placed on each LJ site.
To maintain charge neutrality, a point charge of +0.964e is placed
at the center of mass (COM) of the N2 molecule. The cross interaction parameters are obtained by the Lorentz-Berthelot combining rules. All the size and energy parameters of adsorbates and
adsorbent are given in Table 1.
GCMC simulations16 are carried out to investigate the gas adsorptions inside SWNT and in the interstices between SWNTs.
The SWNT bundles are frozen as a hexagonal structure with the
tube-tube separation of 0.4 nm,17 as shown in Fig.1. This separation distance of 0.4 nm is determined by the experimental data
and also used in the Kowalczykʹs simulations.17 All the SWNTs
are cut off at the z direction with a length of 2.705 nm. The
numbers of unit cells for (18,18) and (20,20) SWNTs are 1×1×1,
while for others SWNT bundles they are 2×1×1. The structural
properties of SWNTs are given in Table 2.
For all the materials, the periodic boundary conditions are
applied in three dimensions (x, y, and z directions). The spherical
cutoff of 1.2 nm is used to calculate the intermolecular LJ interactions without long-range corrections. The electrostatic interaction between fluids is handled by the Wolf spherically truncated
method18 and the cutoff is also set to 1.2 nm. To accelerate the
simulations, the LJ interactions between adsorbates and SWNTs
are interpolated from a pretabulated energy map with the grid
spacing of 0.02 nm. A total number of 2×107 configurations are
generated at every state. The first 107 configurations are discarded
to guarantee equilibration, whereas the remained ones are divided
into 20 blocks for ensemble average. For all the species, the
GCMC procedure will move randomly in translation, rotation,
insertion, and deletion. Instead of using the chemical potential, the
normal move acceptance probabilities are transformed in the
forms of the component fugacity of bulk phase, which is calcu-
2002
Vol.30
Acta Phys. -Chim. Sin. 2014
Table 1 Lennard-Jones and coulombic potential parameters for
C6H6, N2, O2, and SWNT
Parameter
C6 H 6
N2
Diameter/nm
Unit cell
a/nm
b/nm
ρ/(g∙cm-3)
V/(cm3∙g-1)
φ
0.1715
(8,8)
1.085
1.485
2.572
1.36
0.33
0.45
q(z=0.0 nm)/e
+8.130
(10,10)
1.356
1.756
3.041
1.21
0.42
0.51
q(z=±0.04 nm)/e
-4.065
(12,12)
1.627
2.027
3.511
1.09
0.51
0.55
σCH/nm
0.3361
(15,15)
2.034
2.424
4.198
0.96
0.64
0.61
(18,18)
2.441
2.841
4.921
0.84
0.79
0.66
(20,20)
2.710
3.112
5.390
0.77
0.88
0.68
(εCH/kB)/K
75.6
r(N―N)/nm
0.11
q(z=0.0 nm)/e
+0.964
q(z=±0.055 nm)/e
-0.482
(εN/kB)/K
SWNT
SWNT
r(CH―CH)/nm
σN/nm
O2
Value
Table 2 Structural properties for SWNT bundles
0.331
36.0
r(O―O)/nm
0.1208
σO/nm
0.3006
(εO/kB)/K
48.0
σC/nm
0.34
(εC/kB)/K
28.0
r is the bond length of the molecules, q is the partial charge of the interaction
sites, σ and ε are the size and energy parameters of
Lennard-Jones potential, respectively.
lated by Peng-Robinson equation of state.
The isosteric heat (qst) that reflects the affinity between adsorbent and adsorbates is approximated by19
q st ≈ RT - æ ∂U ö
(1)
è ∂N øT,V
where R and T are the universal gas constant and temperature, and
U and N are the total adsorbed energy and number of fluid particles, respectively.
Adsorption selectivity of component i over j in a mixture is
defined as20
æ x öæ y j ö
Si/j = ç i ÷ç ÷
(2)
è x j øè yi ø
where x, y are the mole fractions of component for the adsorbed
and bulk phase, respectively.
The orientation order parameter6 of molecules with respect to
the pore axis is given by a function of the distance r from the
simulation box center to the molecular mass center
S z (r) = 3 cos2 β - 1
(3)
2
2
where the brackets in the equation above denote an average over
Fig.1 A hexagonally-arranged model of SWNT materials
The area within dashed lines denotes a unit-cell used in GCMC simulation, where
D and g denote the pore diameter of nanotubes and the separation
between nanotubes, respectively.
ρ is the materialʹs density, V is the pore volume, and φ is the porosity defined as
the ratio of the pore volume of adsorbent accessible to gas molecules to the
adsorbent volume. The pore volume is calculated by a Monte Carlo integration
with the reentrant surface definition,3 where the argon molecule with a
size of 0.34 nm was used as a probe.
all of the adsorbed molecules; for C6H6 molecules β is the angle
between the normal vector to the benzene ring plane and the pore
axis, and Sz≈-0.5, 1 denote that the molecules are parallel and
perpendicular to the pore axis, respectively; for N2 and O2 molecules β is the angle between the linear vector and the pore axis,
and Sz≈-0.5, 1 denote that the molecules are perpendicular and
parallel to the pore axis, respectively; for all the molecules Sz≈0
denotes that the molecules have no particular orientation.
3
Results and discussion
3.1 Adsorption of pure benzene vapor on SWNTs
Fig.2 shows the adsorption isotherms and isosteric heats of pure
benzene in SWNTs at 303 K. From Fig.2(a, b), we can see that the
capillary condensation occurs at high pressures for all the adsorption isotherms. In addition, the smaller the pore diameter, the
lower the pressure at inflection point. Additionally, the saturation
uptake increases with the nanotube diameter, which can be explained by the material structural data. As shown in Table 2, the
pore volume increases with increasing the pore diameter, while
the SWNT density decreases oppositely with increasing the pore
size. Consequently, the SWNTs with large pore diameter can
accommodate more benzene molecules, but with less material
mass. It is clear that the (20,20) SWNT is more suitable to adsorb
pure benzene vapor. From Fig.2(b), we can also observe that the
adsorption isotherms of (15,15), (18,18), and (20,20) SWNTs first
slightly rise up before the sudden jumps of uptakes. This is because the interstices between these SWNTs can provide additional
adsorption positions for benzene molecules. However, the prior
ascending behavior is not found for (8,8), (10,10), and (12,12)
SWNTs, since the interstices in these materials are inaccessible to
the adsorbate molecules (see Fig.2(a)).
It is also found that the isotherm shapes in the nanotubes are
similar to those in the slit pores.21 Moreover, the jumping pressure
is strongly influenced by the pore diameter for changing several
orders of magnitude, which is identical to the slit pores.21 Furthermore, in the MCM-41 material with the average pore diameter
of 3.52 nm,22 the saturated adsorption amount of benzene is approximately equivalent in the pore size of 2.710 nm for (20,20)
nanotube, both approaching to ~10 mmol∙g-1. Fig.2(c) shows the
No.11 PENG Xuan: Molecular Simulations of the Purification of Toxic Benzene Gas on Single-Walled Carbon Nanotubes
2003
the isosteric heats present a continuous rising for (8,8), (10,10),
and (12,12) SWNTs. This remarkable discrepancy can be attributed to the different adsorption positions between two types of
SWNTs, which is also consistent with the discussion of their
adsorption isotherms aforementioned. It is also observed that the
isosteric heats in the nanotubes fluctuate in the range of 50-100
kJ ∙ mol- 1. Coasne et al.24 simulated the benzene adsorption in
MCM-41 with a pore diameter of 3.6 nm at 298 K. Their isosteric
heats vary between 50 and 100 kJ∙mol-1. This range is consistent
with our case in nanotubes. In the works of Song25 and Zeng26 et
al., the isosteric heats in silicalite-1 and NaY zeolite fluctuate in
50-70 and 80-95 kJ∙mol-1, respectively. All these comparisons
indicate that the isosteric heats of benzene adsorption are greatly
dependent on the types of porous materials.
Fig.2 Adsorption isotherms and isosteric heats of pure benzene in
SWNTs with different pore diameters at 303 K
dependence of the isosteric heat on the uptake of benzene at
different pore diameters. Before the adsorptions are saturated,
narrowing pore size will increase the isosteric heat of adsorption,
due to the greater contribution to the isosteric heats from the
fluid-adsorbent interaction at smaller pores. At saturated uptakes,
the fluid-fluid interaction plays a more important role in affecting
the isosteric heats. Yang et al.23 reported the N2/CO2 adsorption in
IRMOF- 10. In their work, the isosteric heat curve exhibits a
sudden jump before adsorption condensation. Furthermore, the
isosteric heats contributed from CO2-CO2 interaction are much
higher than those from CO2-MOF interaction. Clearly, their
findings validate our results of the benzene adsorption in nanotube.
For the (15,15), (18,18), and (20,20) SWNTs, the isosteric heats
exhibit a local maximum near 1 mmol∙g- 1, then decline to the
plateaus which correspond to the inflection point of pressure on
the adsorption isotherms, and finally rise up sharply. By contrast,
3.2 Separation of benzene from air by SWNTs
Generally speaking, the exhaust gas with a benzene concentration of 50-1000 mg∙m-3 belongs to the waste gas containing
a low concentration of benzene. In this work, we specify 300×10-6 as
the typical concentration of benzene in air. Therefore, the mole
composition of C6H6/N2/O2 gas system is 0.0003:0.79:0.2097. The
simulated total pressure is no more than 4.0 MPa to ensure that the
benzene in the bulk phase is still in the gas state, because the
partial pressure of benzene is 1.2 kPa at this gas composition,
which is much smaller than the saturated vapor pressure of 15.9
kPa. Figs.S1- S6 (Supporting Information) show the pairwise
adsorption selectivities of C6H6/O2, C6H6/N2, N2/O2 and the single
component isotherms at different pore diameters. As expected, the
selectivities of C6H6/O2 and C6H6/N2 are several orders of magnitude higher than that of N2/O2, indicating that C6H6 is the most
preferentially adsorbed among these species. Fig.3 shows the
effects of the pressure and pore size on the selectivity of C6H6/(N2+
O2). As we can see, for all the SWNTs, C6H6 reaches saturation
adsorption at relatively low pressures, while the adsorption isotherms of N2 and O2 increase all along with the pressure. Accordingly, the selectivity of C6H6 gradually decreases with the
pressure. Another factor that distinctly affects the adsorption selectivity is the pore diameter. Interestingly, we see that for all the
pressures, the selectivities exhibit a minimum at the pore diameters D=1.5 nm for (12,12) SWNT, and a maximum at D=2.5 nm
for (18,18) SWNT, respectively, which is caused by the different
responses of the component uptake to the pore diameter. As seen
in Fig.3, the benzene adsorption in air can be enhanced by the
pore diameter monotonously. Nevertheless, with the increase of
the pore size, the uptakes of N2 and O2 transfer from the maximum
to the minimum. Consequently, the (18,18) SWNT is the optimal
material for purifying trace benzene, because both selectivity and
capacity are significantly improved.
3.3 Microscopic structure of benzene mixture
adsorption in SWNTs
To further understand the extremes of adsorption isotherms in
Fig.3, we analyze the adsorption structures for (12,12) and (18,18)
SWNTs. Fig.4 shows the local density profiles of mass center and
probability profiles of the ternary mixture. We can see that for (12,
2004
Acta Phys. -Chim. Sin. 2014
Vol.30
Fig.3 Effects of pressure and pore diameter of SWNTs on the adsorption of N2-O2-C6H6 ternary mixtures at 303 K
12) SWNT at 4.0 MPa, the local density profiles of N2 and O2
exhibit three sharp peaks at r=0.6, 1.5, 2.1 nm, corresponding to
the adsorptions in the interstices of SWNTs. Although the interstices is narrow, it can provide efficient adsorption positions for
these small molecules, where the fluid-solid interaction potentials
between the nanotubes overlap each other. However, the adsorption of C6H6 in (12,12) SWNT mainly occurs inside the tubes,
which is validated by the probability profiles. For (18,18) SWNT
at 0.1 MPa, no peaks are observed for the adsorptions of N2 and
O2. By contrast, the local density profile of C6H6 exhibits distinct
peaks at r=0.3, 0.8, 1.6, 2.1 nm, in succession with the adsorptions
that occur in the center of tube, on the inner surface of tube, and
as the nearest and the second nearest layers in interstices. It indicates that the adsorption of C6H6 is favored at low pressures and
large pores where the adsorption separation ability is dominated
by the enthalpy effect and the entropy effect, respectively.
However, a small quantity of N2 and O2 molecules emerge only in
the interstices at a high pressure of 4.0 MPa. The discussion is
also consistent with the adsorption snapshots in Fig.5.
In order to characterize the orientations of adsorbed molecules
with respect to the pore axis of the SWNTs, we calculated the
order parameter Sz(r) as a function of the distance r from the
center of simulation box to the mass center of adsorbate. From
Fig.6, we can see that for both materials the benzene molecules in
the contact layer and interstices give an approximate value of -0.5
for Sz, which shows that these benzene molecules prefer an orientation lying nearly flat on the pore surface. Interestingly, for (18,
18) SWNT, the benzene molecules beyond the contact layer (r=
0.7 nm) show a different orientation, where their benzene rings are
nearly perpendicular to the pore axis. This is because these
molecules tend to recover their bulk properties for the less confinement. Note that the above orientation profiles are also in good
consistence with the molecular dynamic simulations by Bhide and
Yashonath.27 For N2 and O2 molecules, their orientations in both
No.11 PENG Xuan: Molecular Simulations of the Purification of Toxic Benzene Gas on Single-Walled Carbon Nanotubes
Fig.4
2005
Local density profiles (top panels) of molecular mass center as a function of the distance r, and probability profiles (bottom panels)
of molecular mass center from top views for N2-O2-C6H6 adsorption
The left panels are for SWNTs (12,12) at 4.0 MPa and the right panels are for SWNTs (18,18) at 0.1 MPa. The coordination origin is in the center of simulation box.
Fig.5
Snapshots of N2-O2-C6H6 ternary mixtures in SWNTs
(a-c) SWNTs (12,12) at 4.0 MPa, (d-f) SWNTs (18,18) at 0.1 MPa. (c) and (f) are adsorptions in the interstices between tubes, others are inside tubes.
materials are basically identical, i.e., adopting a parallel angle to
the pore axis. All these preferential orientations also can be seen
in the molecular configurations in Fig.5.
3.4
Effects of temperature and concentration on
benzene separation
Fig.7 shows the effect of temperature on the separation of
2006
Fig.6
Acta Phys. -Chim. Sin. 2014
Orientational order parameters Sz(r) of adsorbed N2-O2C6H6 ternary mixtures
(a) SWNTs (12,12) at 4.0 MPa, (b) SWNTs (18,18) at 0.1 MPa
benzene at the SWNTs with different pore sizes. It is well known
that the increase of temperature will reduce the adsorption se-
Fig.7
Vol.30
lectivity, which is confirmed by the adsorption behaviors in (15,
15), (18,18), and (20,20) SWNTs. Furthermore, Fig.7(d) shows
that the uptakes of benzene fall down sharply from the saturated
amount to 1 mmol∙g-1 when the temperature is increased by 50 K.
However, a further release of benzene is quite difficult, since it
requires a large temperature rise of 150 K to completely remove
the residue. Accordingly, the selectivity curve declines in the
whole temperature range, as shown in Fig.7(b). However, for (8,
8), (10,10), and (12,12) SWNTs, the situation is a little different
from the previous case. From Fig.7(a), we see that with the increase of temperature, the selectivity of benzene shows a rise first,
and then drops down. However, the uptakes of benzene are almost
constant in the initial heating process (see Fig.7(c)). Consequently,
this prior ascending of selectivity can only be interpreted by the
sharper decrease of the uptakes of N2 and O2 molecules. By
comparing these two types of SWNTs, we found that the selectivities and uptakes in the larger pores where the confinement
effect is weaker, are more influenced by the temperature. Clearly,
from the viewpoint of adsorbent regeneration, the temperature
swing adsorption (TSA) process could be an appropriate way for
the best carbon nanotube material of (18,18) SWNTs for benzene
purification.
Fig.8 shows the dependence of the selectivity and uptake on the
benzene concentration of bulk phase. From Fig.8(d), we can see
that the uptakes of benzene go up drastically in the large pores.
However, the selectivities of benzene exhibit a slight decrease
because the concentration of benzene in the bulk phase increases
greater than that in the pore phase by several orders of magnitude
Effect of temperature on the adsorption of N2-O2-C6H6 ternary mixtures in SWNTs at 0.1 MPa
No.11 PENG Xuan: Molecular Simulations of the Purification of Toxic Benzene Gas on Single-Walled Carbon Nanotubes
Fig.8
2007
Effect of bulk benzene concentrations on the adsorption of N2-O2-C6H6 ternary mixtures in SWNTs at 303 K and 0.1 MPa
(see Fig.8(b)). The similar behavior can be observed in the small
pores. However, for the (10,10) and (12,12) SWNTs, the uptakes
undergo an increase first and then reach a stable plateau (see Fig.8
(c)). As a result, there appears a maximum of the benzene selectivity in Fig.8(a).
References
(1)
Benzene Poisoning in Chemical Laboratories. J. Chem. Educ.
1929, 6, 513.
(2) Yan, Y. P.; Chen, F. Z.; Shao, Y. X. Tech. Equip. Environmental
Pollut. Control 2000, 1, 76. [颜幼平, 陈凡植, 邵英贤. 环境污
染治理技术与设备, 2000, 1, 76.]
4
Conclusions
By using grand canonical Monte Carlo simulation technique,
we have systematically investigated the purification of benzene
from air in SWNTs. We find that a large pore size of (20,20)
SWNT is more efficient to adsorb pure benzene vapor. However,
for the removal of trace benzene in air, (18,18) SWNT is the best
separation material, because the maximum of selectivity (>106) is
obtained at 0.1 MPa. The microscopic structures such as molecular configurations, local density profiles, and probability profiles
are further analyzed for the N2-O2-C6H6 mixture. We observe that
(18,18) SWNT only adsorbs the benzene molecules. The orientation order parameters indicate that benzene molecules nearly flat
on the pore surface. The effects of temperature and concentration
of bulk benzene on the selectivity of benzene are finally studied.
Our simulations reveal that the carbon nanotubes could be excellent adsorbents in effectively purifying the toxic benzene impurity in air.
(3)
Peng, X.; Cao, D. AIChE J. 2013, 59, 2928. doi: 10.1002/aic.
v59.8
(4)
Peng, X.; Cheng, X.; Cao, D. J. Mater. Chem. 2011, 21, 11259.
doi: 10.1039/c1jm10264h
(5)
Do, D. D.; Do, H. D. Langmuir 2006, 22, 1121. doi: 10.1021/
la052545i
(6)
Coasne, B.; Fourkas, J. T.; Normale, E. J. Phys. Chem. C 2011,
115, 15471. doi: 10.1021/jp203831q
(7)
Jousse, F.; Auerbach, S. M.; Vercauteren, D. P. J. Phys. Chem. B
2000, 104, 2360. doi: 10.1021/jp9935642
(8) Amirjalayer, S.; Tafipolsky, M.; Schmid, R. Angew. Chem. Int.
Edit. 2007, 46, 463.
(9)
(10)
Iijima, S. Nature 1991, 354, 56. doi: 10.1038/354056a0
Cinke, M.; Li, J.; Bauschlicher, C. W.; Ricca, A.; Meyyappan,
M. Chem. Phys. Lett. 2003, 376, 761. doi: 10.1016/S0009-2614
(03)01124-2
(11) Wang, W.; Peng, X.; Cao, D. Environ. Sci. Technol. 2011, 45,
Supporting Information: The adsorption selectivities and
single-component isotherms of ternary mixtures in six SWNT
bundles have been included. This information is available free of
charge via the internet at http://www.whxb.pku.edu.cn.
4832. doi: 10.1021/es1043672
(12)
Huang, L.; Zhang, L.; Shao, Q.; Lu, L.; Lu, X.; Jiang, S.; Shen,
W. J. Phys. Chem. C 2007, 111, 11912. doi: 10.1021/jp067226u
(13)
Bonnaud, P.; Nieto-Draghi, C.; Ungerer, P. J. Phys. Chem. B
2008
Acta Phys. -Chim. Sin. 2014
2007, 111, 3730. doi: 10.1021/jp067695w
(14)
Gubbins, K. Phys. Chem. Chem. Phys. 2011, 13, 3748. doi:
Perng, B.; Sasaki, S.; Ladanyi, B. M.; Everitt, K. F.; Skinner, J.
L. Chem. Phys. Lett. 2001, 348, 491. doi: 10.1016/S0009-2614
10.1039/c0cp02205e
(22)
(01)01152-6
(15)
Potoff, J. J.; Siepmann, J. I. AIChE J. 2001, 47, 1676.
(16)
Frenkel, D.; Berend, S. Understanding Molecular Simulation;
Academic Press: London, 2002.
(17)
(23) Yang, Q.; Ma, L.; Zhong, C.; An, X.; Liu, D. J. Phys. Chem. C
2011, 115, 2790. doi: 10.1021/jp1101835
(24)
007-9051-3
(25)
Song, L.; Sun, Z.; Ban, H.; Dai, M.; Rees, L. Adsorption 2005,
(26)
Zeng, Y.; Ju, S.; Xing, W.; Chen, C. Ind. Eng. Chem. Res. 2007,
(27)
Bhide, S.; Yashonath. S. J. Phys. Chem. B 2000, 104, 11977.
Phys. 1999, 110, 8254. doi: 10.1063/1.478738
Peng, X.; Cao, D.; Wang, W. Ind. Eng. Chem. Res. 2010, 49,
(20)
Peng, X.; Wang, W.; Xue, R.; Shen, Z. AIChE J. 2006, 52, 994.
(21)
Coasne, B.; Alba-Simionesco, C.; Audonnet, F.; Dosseh, G.;
Coasne, B.; Alba-Simionesco, C.; Audonnet, F.; Dosseh, G.;
Gubbins, K. Adsorption 2007, 13, 485. doi: 10.1007/s10450-
doi: 10.1021/es071306+
(19)
Nguyen, D.; Do, D. J. Phys. Chem. B 2000, 104, 11435. doi:
10.1021/jp0007282
Kowalczyk, P.; Holyst, R. Environ. Sci. Technol. 2008, 42, 2931.
(18) Wolf, D.; Keblinski, P.; Phillpot, S. R.; Eggebrecht, J. J. Chem.
Vol.30
11, 325.
8787. doi: 10.1021/ie1010433
46, 242. doi: 10.1021/ie060118+
doi: 10.1021/jp002626h
Supplementary Information for Acta Phys. -Chim. Sin. 2014, 30 (11), 2000-2008
doi: 10.3866/PKU.WHXB201408291
单壁碳纳米管上毒性苯气体净化的分子模拟
彭 璇*
(北京化工大学信息科学与技术学院,
北京 100029)
Molecular Simulations of the Purification of Toxic Benzene Gas on
Single-Walled Carbon Nanotubes
PENG Xuan*
(College of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, P.
R. China)
∗
Corresponding author. Email: [email protected], [email protected]; Tel: +86-10-64430917.
S1
10
Se lec tivit y
10
10
10
10
9
C H /O
(a )
6
6
2
C 6H6 /N 2
O 2/N 2
C 6H6 /( N2+O 2)
7
5
3
1
-1
10
0
1
2
3
4
Pre ss ur e (MPa )
3.0
( b)
Up ta ke ( mmo l/g )
2.5
2.0
1.5
C 6H6
O2
N2
1.0
0.5
0.0
0
1
2
3
4
Pre ss ur e (MPa )
Fig.S1 Adsorption selectivities and isotherms of N2‐O2‐C6H6 ternary mixtures (yN2=0.79 and yC6H6=0.0003) in (08,08) SWNTs at 303 K S2
10
Se lec tivit y
10
10
10
10
10
9
C H /O
(a)
6
6
2
C 6H6/N 2
O 2/N2
C 6H6/(N 2+O 2)
7
5
3
1
-1
0
1
2
3
4
Pre ss ur e (MPa )
5
(b )
Up ta ke (mmo l/g )
4
3
CH
6
6
O
2
2
N2
1
0
0
1
2
3
4
Pr es su re ( M Pa)
Fig.S2 Adsorption selectivities and isotherms of N2‐O2‐C6H6 ternary mixtures (yN2=0.79 and yC6H6=0.0003) in (10,10) SWNTs at 303 K S3
10
Se lec tivit y
10
10
10
10
10
9
C H /O
(a)
6
6
2
C 6H6/N 2
O 2/N2
C 6H6/(N 2+O 2)
7
5
3
1
-1
0
1
2
3
4
Pre ss ur e (MPa )
6
(b )
Up ta ke (mmo l/g )
5
4
3
CH
6
6
O
2
N2
2
1
0
0
1
2
3
4
Pr es su re ( M Pa)
Fig.S3 Adsorption selectivities and isotherms of N2‐O2‐C6H6 ternary mixtures (yN2=0.79 and yC6H6=0.0003) in (12,12) SWNTs at 303 K S4
10
Se lec tivit y
10
10
10
10
10
9
C H /O
(a)
6
6
2
C 6H6/N 2
O 2/N2
C 6H6/(N 2+O 2)
7
5
3
1
-1
0
1
2
3
4
Pre ss ur e (MPa )
10
(b )
Up ta ke (mmo l/g )
8
6
CH
6
6
O
4
2
N2
2
0
0
1
2
3
4
Pr es su re ( M Pa)
Fig.S4 Adsorption selectivities and isotherms of N2‐O2‐C6H6 ternary mixtures (yN2=0.79 and yC6H6=0.0003) in (15,15) SWNTs at 303 K S5
10
9
10
7
10
5
10
3
10
1
(a )
C H /O
6
6
2
C H /N
Se lect ivity
6
10
6
2
O2/N 2
C6H 6/(N2+O 2)
-1
0
1
2
3
4
Pr es su re ( MPa)
12
(b )
Up ta ke (mmo l/g )
10
8
6
CH
6
6
O
4
N2
2
2
0
0
1
2
3
4
Pr es su re ( M Pa)
Fig.S5 Adsorption selectivities and isotherms of N2‐O2‐C6H6 ternary mixtures (yN2=0.79 and yC6H6=0.0003) in (18,18) SWNTs at 303 K S6
10
9
10
7
10
5
10
3
10
1
(a )
C H /O
6
6
2
C H /N
Se lect ivity
6
10
6
2
O2/N 2
C6H 6/(N2+O 2)
-1
0
1
2
3
4
Pr es su re ( MPa)
14
(b )
Up ta ke (mmo l/g )
12
10
8
CH
6
6
O
6
2
N2
4
2
0
0
1
2
3
4
Pr es su re ( M Pa)
Fig.S6 Adsorption selectivities and isotherms of N2‐O2‐C6H6 ternary mixtures (yN2=0.79 and yC6H6=0.0003) in (20,20) SWNTs at 303 K S7